Main Edition
US
FR
We have nothing to fear but fear itself. Apart from the deadly pandemic. Photo: Marco Cantile/LightRocket/ Getty
We all rely on the news to give us information about the world. That information lets us make decisions: whether it’s safe to fly to Spain, whether red wine causes cancer, whether we’re likely to lose our job for tweeting something. We use the media to help us understand the risks that surround us. The trouble is that the media is uniquely bad at telling us about those risks.
Let’s imagine that you want to know whether water causes cancer. You do a simple study: you go to your local hospital, look at all the people who’ve got cancer, and check whether they had recently drunk some water. You find that almost all of them had, so you conclude that water is a carcinogen.
Very obviously, this is not a good way of doing the study. Since almost everyone drinks water, you’d find that almost everyone who had cancer had drunk water, even if drinking water in no way increased your chances of getting cancer. What you need to do, instead, is compare people who did get cancer with people who didn’t, and see whether one group is more likely to have drunk water lately.
(As it happens, the scientific consensus at the moment leans towards the idea that drinking water is a good idea.)
This is an example of a mistake called “selecting on the dependent variable”. In science, the dependent variable is the thing you’re trying to measure; the independent variable is the thing you change. So in a drug trial, the independent variable might be the dose of the drug; the dependent variable might be whether or not the patient survives. If you only looked at the patients who survived, the results of your drug trial would not be very useful.
In a subtle way, though, this is what the media does all the time. Like the water-causing-death study, it selects on the dependent variable: and for the media, the dependent variable is whether it’s interesting.
Join the discussion
Join like minded readers that support our journalism by becoming a paid subscriber
To join the discussion in the comments, become a paid subscriber.
Join like minded readers that support our journalism, read unlimited articles and enjoy other subscriber-only benefits.
SubscribeThe media appear unwilling to listen. We know that the Imperial model, which projected 500,000 as it’s worst case, is seriously over stated yet that number still appears regularly on both television, print and pixel. I still see the 45,000 PHE death figure quoted daily, usually as a stick to beat someone with, yet this number is suspect in the extreme and is now being investigated.
The media went all out to in calling Covid 19 a pandemic. They have ignored risk and still ignore the huge cost, both in lives and in economics, that the lockdown has imposed on this country.
The question is will the media be held to account for it’s woeful performance since the virus was identified?
I’m afraid not. The media will simply roll on, as it has always done.
45,500 deaths 2% (900) due to SARS2 is never reported by BBC,ch4,ITV,cnn,or Sky 24/7 hysteria and ”Fake Science” also on Climate We are asked to believe Weather Climate Change experts who cannot predict 10 days ahead ”,Forecast”100 years hence; polar bears ‘Could” disappear unlikely unless penguins ,Fish die out..
Quit right, Climate models need serious review but I guarantee no one will look at them for fear of what they might find out!
Er, no, but you have given an example of how issues can be conflated. Weather and climate are two completely different things, involving completely different processes, data, and timescales.
Agree 100%. Not only do people not understand the difference between causation and correlation, they also have no idea how many people die every day, and no appreciation of risks we undertake unconsciously.
AS the picture changes and we start to drop down the ranks of case and death numbers, (And certainly when excess deaths are published) will the media hold it’s hand up and say ‘we got this wrong’?. I doubt it.
We can now see case numbers rise dramatically in countries which were judged by the media to have ‘got it right’ 2 months ago. How long until they pivot?
If their commentary about the quarantining of returnees from Spain is anything to go by, we will have a long wait.
Exactly. Germany was held up as the exemplar of getting it right.
Now, ‘Coronavirus: German officials ‘very concerned’ by rising cases’
https://www.bbc.co.uk/news/…
Politicians more than the media mislead us by focusing on the IC and other, e.g. LSHTM, model worst case scenarios. Covid-19 is a pandemic and lockdown was imposed to avoid the higher costs in lives and to the economy of not imposing it.
It’s very easy to knock the Imperial 500,000 figure, but I’ve never really understood why anyone would bother. It was always a theoretical worst case that the modellers themselves did not expect to happen; “In the (unlikely) absence of any control measures or spontaneous changes in individual behaviour, we would expect a peak in mortality (daily deaths) to occur after approximately 3 months” of 510k deaths and 81% of the population having had the virus.
While your point that it was a worst case is true, it was never plausible.
There are no infectious disease outbreaks which infect 80%+ of a population. Spanish flu infected around 30% of the world’s population.
The larger discussion point is whether lockdown saved many lives, few or none. The argument is long forgotten but the answer was claimed to be dependent on the extent to which demand exceeded NHS capacity minus the degree to which lockdown might flatten the peak.
Personally I think it’s moot. We were fortunate not to breach capacity & non human factors caused the infection rate to decline before lockdown.
“There are no infectious disease outbreaks which infect 80%+ of a population”
While the common cold is much more complicated, probably about that proportion of the population get some form of cold in a year. The Black Death killed about 60% of the population of some cities and 80% of one, and European diseases, such as smallpox, may have killed 90% of the population in some part of the Americas — they must therefore have infected those proportions at a minimum. Bear in mind that we are talking about a novel disease invading a population which cannot be assumed to possess pre-existing immunity.
Not a fair comparison, we now know things like to wash our hands, stay home when sick, eat healthy foods (like fully cooked meats). We’ve learned that covid does not infect 80% of the population now. We based these lockdowns on these doomsday predictions, and we now know that suicides and drug overdose deaths are higher than covid deaths. People have lost their livelihoods and retirements.
People need to bother because Imperial and other institutions have a history of lassitude bording on negligence by displaying theoretical models as applied analysis. Eg F&M disease. If you pulled a stunt like that in pharmacology or engineering at best you’d be fired and at worst jailed. If their claims were preceeded by the caveat “here’s an opinion from a biased academic with Donald Trump’s sense of restraint and a financial interest in making these claims” then they would rightly be ridiculed and then ignored. As it is they are given the same weight as testable, replicable evidence based studies – which is why they are a problem which affects so many lives.
The Imperial models have issues not least there long history of inflating results from Foot and Mouth through to Sars. We also need to look directly at the working of the model as it appears to be so ad hoc that no one has been able to replicate the findings and Ferguson’s paper has not been peer reviewed.
But the real issue with the 500K figure has been the way it has been used in the media as if it is chiselled in stone. They never critically analysised the report just hyped it up!
> I still see the 45,000 PHE death figure quoted daily, usually as a stick to beat someone with, yet this number is suspect in the extreme and is now being investigated.
The author’s own Twitter shows that this take is wrong (and is an example of the sort of thing the article is talking about, since the whole “PHE is wrong” thing is hyped by certain sections of the media): https://twitter.com/TomChiv…
Looking at excess death figures shows the PHE figure is likely to be an underestimate.
“Looking at excess death figures shows the PHE figure is likely to be an underestimate.” – Why? We need to look at the excess death figures yes, but excess deaths does not mean deaths from Covid 19. Cancer patients are dying, heart patients are dying, stroke patients are dying etc because the NHS are not treating them to leave space for Covid 19 patients. I know a number of people who where having cancer treatment in February with expect good outcomes who had their treatment stopped in March and have now been put on palliative care due to the advance of the cancer.
This is going to be one of the biggest post lockdown issues, the people who died of other causes. It is vital we do not assume all excess deaths are from Covid 19, they most certainly are not!
Are we agreed that Tom Chivers is correct to say that the PHE counting any positive test followed by a death can only have inflated the figures by a few tens of deaths per day, so made little difference at the peak of the epidemic?
> Cancer patients are dying, heart patients are dying, stroke patients are dying etc because the NHS are not treating them to leave space for Covid 19 patients.
People have died and will die because they were unable to get treatment or too scared to go to a hospital. But I see no reason why excess deaths from these causes should peak and then decline a few weeks after lockdown was imposed in earnest (https://www.reddit.com/gall… has the graph). Lockdown reduces novel coravirus spread and, a few weeks later, COVID19 deaths (as well as reducing other diseases affected by social distancing).
Attendance at hospitals for cardiac conditions dropped dramatically at the point where lockdown was imposed and gradually recovered, but we can do a similar thing to excess deaths and look at “missing” patients, and find that even if all the missing cardiac patients died, this doesn’t account for the excess death figures from March, April and May (https://assets.publishing.s… graph 12 “Cardiac”: there are about 9 weeks of missing patients, the fall is 500 per day down to 250 per day and back again, call it a triangle and that’s about 7875 deaths).
> I know a number of people who where having cancer treatment in February with expect good outcomes who had their treatment stopped in March and have now been put on palliative care due to the advance of the cancer.
I do expect cancer patients will die earlier because they couldn’t get care, but again, these deaths are mostly in the future and don’t explain the size of the excess death peak that occurred around Easter this year.
Well, if the cancer care was stopped due to covid-19 rules or cultural effects, then the excess deaths are an indirect result of the covid-19 pandemic.
It would be useful to know what proportion of the gap between identified covid deaths and excess deaths were directly caused by the virus. But it will not be 100% of the difference.
The cancer care was stopped because of our reaction to Covid 19, not because of it. The cure (lockdown) is wors e than the disease!
I read this article, indeed most of your articles, with engaged interest, nodding along in sympathy to the points made, but I’m the wrong audience, I already critique every number and statistic I read, the people who really need to read this stuff aren’t reading it! The challenge is how we get this stuff into the minds of those who really need it… I wish I knew the answer to that.
Good point. I get the same feeling about ‘More or Less’ on R4.
It’s consistently good, but I have the feeling that its listeners are the people least easily fooled. It’s everybody else who needs it more.
R4 consistently good ..how do you know ? It’s just anther dubious squawk programme . You’re being arrogant in thinking and writing your post DG
Some of the rubbish that comes out of a couple of it’s listeners whom I know fairly well is risible. Most of the time sadly , it seems that they are incapable of thinking things through for themselves .
“More or Less” on R4 is the best fact checker I know. Their Covid-19 series is excellent. Avail on bbc sounds.
This is undoubtedly about the biggest problem in the World…”The Media controls the minds of the people” Jim Morrison The Doors Rock Group. Perhaps “propaganda” & “hoodwinking” the general Public can be for good reasons….but seems to me, as with business advertising, a price needs to be exalted upon “false & misleading” Media storeys & also those of Politicians… I imagine Iraq invasion might be an example….On the bright side orgs like UnHerd are a reason to be cheerful…IF state & private “censorship” (unfortunately too often now branded “Moderation” at The Conversation) doesn’t ruin what can be a brilliant “internet revolution” for the long neglected folks of this magnificent Planet!
Schools. We should teach children critical thinking and about how thinking paradigms can mislead (and how they also deliver good thinking.)
The book is called “Good Thinking and Bad”. it is on amazon.
Full disclosure- I’m the author.
Actually I think articles like this are the best for sharing with anyone! They are light, objective, stick to facts and reliable sources, and avoid entering conspiracy spirals. It’s critical thinking without being hard on those who think differently or trying to forcefully/emotionally push opinions – which is usually what gets people to close their minds and ears.
I can’t believe anyone still believes that these colossal gaffes are simply “mistakes”. Please show me one MSM article that accurately portrayed the common sense approach to “protect the high risk while the rest of us achieve herd immunity”. (Opinion pieces don’t count, that’s the only place they were published, if they were published.)
Good article and much relieved that a manual for the media on the use of numbers and their interpretation is being produced. Hopefully it will also be read by the wider public too. My grasp of statistics is limited to having studied for a term or two as part of the foundation of a social science degree fifty years ago. Yet even I have been become hopping mad about the way unbalanced and misleading media articles and especially headlines are produced on a regular basis by the so-called mainstream media. The commentary on Covid has simply made a bad situation worse. Sometimes the headlines are not even borne out by the articles to which they refer. Best of luck in getting some commonsense and numeracy into the system.
I think the problem is that many (younger?) journalists nowadays are activists. Their goal is not to report, it is more to lead us to a better place. Worse, many seem convinced that their readers are less smart than they are. Given the raw information might make them chose the wrong politicians! Now it is true that most journalists have a very bad understanding of numbers, and especially percentages. Generally, if stupidity is a possible explanation it is the best explanation. However, in the case of journalists their low knowledge of numbers let them pick numbers from others that sounds most alarming. And once they are out there, the tend to repeat them with a very high R.
But they are stupid as well.
They also exhibit the Dunning”“Kruger effect.
It’s the same with teaching. It’s no longer enough to just teach, you have to be an advocate for social justice too. Unfortunately too many teachers already have a savior complex and so social justice ties in very well with this view of themselves. They see themselves as soldiers of enlightenment mucking down in the ‘trenches’ rescuing the poor souls in their custody from the influence of their ‘misguided’ parents.
Thankfully, not all teachers are like this, however this mentality is also pervasive among school and college administrators too. If a student fails a class it is often ascribed to ‘bad’ teaching rather than student responsibility. Many of the professional development sessions I am made to attend are about ‘ensuring student success’. The bottom line of this is, of course, funding. Colleges in the US are funded on performance measures – student graduation rates being one of them. All the DIE stuff is fluff that obscures this while at the same time giving administrators that warm fuzzy space cadet glow of social justice.
Going back to journalism – you can’t be both an activist and a journalist. Journalists are supposed to report objectively. If you have an agenda (no matter how seemingly noble) you compromise that objectivity and eventually lose the trust of the public. This is currently the case with much of legacy media. They misappropriate the idealism of the young and manipulate their naive sense of fairness and justice.
Maybe not exactly a better place, more a place of their crusading choosing.
99.95% of all men who drop down dead in the street are wearing underpants! Clearly we all need to go commando.
Breaking news – huge spike of common cold cases in London. Mayor of London says new lockdown imminent. Reports are emerging of increased deaths in care homes related to this spike – increased cases of pneumonia and sepsis. PHE has issued guidance to hospitals and GPs to record deaths as common cold deaths if clinician is unclear of exact cause of death. Government announces increased testing for common cold and also warns of a second wave. Mask wearing will become mandatory for 24 hours a day. Gloves will also have to worn.
Breaking news – to help rescue the economy the Government to announce measures to encourage to people to go out shopping and using leisure facilities. The PM says we need to be confident in leaving our homes.
I’d say the problem with the media is the people work in it aren’t particularly adept at handling information. They just don’t have the education for it. In my life I’ve dealt mostly with things border on engineering. Where numbers matter. When provided with information I’ll tend to do a reality check, based on something I’ve experienced. If this number & that number are
correct, does the outcome resemble what I’m being told? It’s not necessary to be that accurate. Quick mental arithmetic. Very often the errors an order of magnitude out. No, there weren’t 150,000 people demonstrating in that street. The street’s not big enough to hold more than 15,000, packed shoulder to shoulder. If life expectancy is 85, how many deaths a year should you expect to see in a population of 60 million? Are Covid fatalities reported a significant figure or little more than a rounding error on a much bigger number? Sorry, but from what I’ve experienced of the arts graduates infest the media, they’d struggle to price the contents of a supermarket basket without resource to a calculator. There were couple of them, in an infamous clip from a US news show, agreeing sagely with a claim that would have meant the election spend of a political candidate translated into a million dollars per voter.
Are we really expected to believe that most journalists are too daft to realise that they are misrepresenting the data ?
I’d love to see a controlled study that backs this dubious claim up …
The author is being over generous.
Excellent article Tom, although I don’t believe that most media organisations, including the BBC, will care (to their eventual cost) about your suggestion. Their sole prerogative is to gain as much profit as possible (naturally enough) and they discovered early on that by always embroidering and warping the truth and banging out overblown headlines, they increase profits with enough left over to hire lawyers and pay off complainants. We know this, so people should treat everything they say, by default, as hogwash, and use it only to find the actual truth (and tone) via reliable news organisations – those that employ the methods that Tom lays out – and via trusted primary sources.
This should be taught in primary schools.
I have so educated my kids along these lines. People are also not taught to ask “why are these stories being presented to me today?” (As well as your point on in what way). It’s a thought provoking question!
Michael it’s a good thing to also instill in your kids minds ” Be careful what you put in your mind , because if you aren’t someone else will . Then over time explain it and how gullible people are …
PS .If you say Gullible slowly it sounds like oranges
Gullible is the word!
Couldn’t agree more. But, where do we find reliable news sources? I gave up all television news drama many years ago. I haven’t bought a newspaper in 25 years. I endure some of the Today program though this is just a cheap sniping shop against the government and has become a platform for carefully scripted whinging and whining nobodies snivelling about how their life has changed. Even worse the number of ‘former’ somebodies with an axe to grind backed up by ‘experts’ from never heard of organisations.
The internet mmm!… Google will block attempts to access potentially interesting outlets at the same time bombarding American content at you.
So where are interesting reliable news outlets, Please inform.
Thank you.
GF notes “Google will block attempts to access potentially interesting outlets “….my suspicion is it might have been “told” or had it “suggested” by “Big Brother” to “play the game” & “moderate” or else be targeted…with dire consequences, …I think Google used to have superb information in first few pages of a search…now one might need to look for quite a long time…also, links have definitely been put down the “Memory hole”. Seems to me sadly Google, Apple & Microsoft et al might NEED to do what they are told to…
Yup. Sensation sells. Good news doesn’t.
Excellent article. I think you are too generous to journalists. Misrepresentation of numbers, and therefore, facts has been going on for ever. Ben Goldacre has put this very well on medical research reporting. It’s hard to believe that journos are not looking for a story and are fully prepared to say “damn the detail”.
Its also hard not to believe that our politicians are not deliberately doing the same. I think the BorisCummings show have become very good at doing this.
Its the modus operandi of popularist rule.
The mistakes may be innocent, but since journalists are always looking for “the angle” they’ll never be neutral, they’ll always mislead.
This is a problem caused by writing ‘stories’ about the real world.
It is for the most part noticeable in local press, just how, dare I say, boring they can appear. Especially when on line having just read an influencer piece or major exposé in the nationals. However, they are simply reporting what happened what they heard and what they saw. And that is as it should be.
Kellyanne Conway was the first woman ever to be the campaign manager for one of the main party candidates in a US presidential election, and of course, won that election. (Tanya wears her feminism on her sleeve, but she has never applauded Kellyanne for that impressive achievement, and I suspect she never will.) Kellyanne said that even in the darkest days of that campaign the Trump forces never gave up hope, because their polling showed that there was so much revulsion in a big part of the American public against Hillary, most Americans would never vote for her, no matter what happened during the campaign.
Tanya keeps playing the victim card in this piece, and we are asked to believe that Hillary’s presidential campaign was torpedoed by nasty American misogyny, and nothing else. It is too silly a picture for words. For a picture that bears some semblance to reality I would recommend Dianah Johnstone’s “Queen of Chaos: The Misadventures of Hillary Clinton”. It’s only weakness is that it was written in August 2015, and so is a little out of date. Hillary is so rebarbative that even a life-long socialist like Johnstone, faced with a choice between her and Donald Trump, much preferred the Donald.
Wrong sort of woman. Cf. Priti Patel and Kemi Badenoch, the wrong sort of BAME women.
The issue is not that she is a woman but that she is wholly corrupt. There is an interesting debate to be had whether she is more corrupt than her husband. I think she is certainly more intelligent than Bill, which leads me to suspect she was the brains behind a lot of the Clinton family corruption.
In 2016 the least worse candidate won – it had nothing to do with what was or was not dangling between their respective legs. A lot of observers from afar can’t believe the yanks ever elected someone as grotesque and the Donald and people like Tayna like to blame it on misogyny. The same will happen in 2020 as the American people yet again have to decide between 2 candidates wholly unsuitable for the most powerful office in the world. My money is on Trump sneaking through again.
There must be plenty of honest and competent women in the Democratic party (just not the one who tried to claim a native American heritage by virtue of being 1/1024th native) and I am sure had they selected one instead of the male turkeys who ran for the nomination, she would be able to wipe the floor with Trump, but after Hillary I think they are just too scared to do it.
My favourite quote from Hillary Clinton comes from a 1998 speech she gave on domestic violence:’Women have always been the primary victims of war. Women lose their husbands, their fathers, their sons in combat.’
Gold standard lack of self-awareness in two sentences.
There have been many really effective women Governors and Cabinet members in the US but not her.
Another example is the very common, ‘ x many people died after testing positive for covid 19’. The implication is that they died FROM C19 when in fact they may have died from entirely unrelated causes. The BBC still uses this stock phrase in a deliberate attempt to play on people’s fear.
Very true. For example much fuss that more tickets have been handed out to Black people for lockdown breaches than their proportion of the population- BBC immediately suggested racism . But unless you know the number of actual offences in each group you cannot tell if it deviated from the mean . The Media – particularly the BBC -needs to use numerical data properly.
They’re not interested in using data properly. They’re interested in promoting a particular narrative, e.g. “dozens of police injured in mostly peaceful protests,” or “Protesters in California set fire to a courthouse, damaged a police station and assaulted officers after a peaceful demonstration intensified.”
A peaceful protest intensified. Seriously. That’s what was said:
https://mobile.twitter.com/…
The media right now are promoting the CV19 scaremongering, talking about “cases” increasing, but never clarifying what a “case” is exactly. Is it a hospital admission, someone showing symptoms, or just a positive test result? They’re saying nothing about recovery rates, successful treatments that are now available and being used.
It’s just “Shut up. Don’t ask awkward questions. Wear your mask or else. Wait for a vaccine. Then you might be allowed to have your old life back, but don’t bet on it, ‘coz of the New Normal and all that. Oh, and let’s ‘Build Back Better.'” Repeat ad nauseam.
Do you get the impression that the types that go to Davos and other meetings that we never hear about, have taken the opportunity to hide the second failure of global capitalism, in ten years, by agreeing to similarly ‘damage’ each economy and to blame the virus/us?
The nerve of capitalist organisations and right leaning low tax government in going down the socialist self defence route is astonishing when the likes of Mrs Thatcher insisted that society as we understood it did not exist. Yet they are already lining society up for tax increases and doubtless, lowering of standards and services.
As a former colleague in educational research was wont to observe:
“Coincidence is not correlation;
“Correlation is not causation.”
I used to have my statistics classes chant “Correlation does not imply causation” while they settled into their seats at the start of a lesson.
This is true, as far as it goes. It can sometimes mean causation, though. More importantly, finding a correlation doesn’t mean there’s nothing to explain, but the “correlation does not imply causation” mantra is often presented as if it did.
If there is a substantial correlation, and a reasonable theoretical basis of causation, and no reasonably credible alternative theory, then the observation of that correlation can be taken to be strong evidence of that causation. For instance if there were a sharp increase of brain cancer in line with mobile phone usage, you would have to be incredibly dense, or have become very well brainwashed in Philip’s lectures, to not recognise the strong implication that the phones have caused the cancer. The implication is all the stronger if the theory preceded the observation of the correlation.
bleah. Drivelly. Wandering. Self-sorrowing.
So sad for poor Hillary, wife of a two-term President, Secretary of State, Senator, with $75,000,000+.
Poor, poor Hillary. What a sad victim. I’m crying. Boo hoo.
Reading Matt Ridley’s book ‘How Innovation Works’ he points out that the pandemis was not taken seriously to begin with because there have so many over-hyped pandemics or ‘wolf cry’ in the past.
Yet the screaming hords still think the opposite.
May I enquire if English is your first language?
It is but I have to type blind as Unherd scrolls up when I press a key. I can’t be bothered to feed idiots when they don’t want to listen
I see you have resorted to the abuse pretty quick as per usual. Here’s a study I would be interested : The number of people who have worked out that the way to get past the Un herd jumping edit window problem is to write your response in a text editor then simply copy it across and split that by the number of people who believe in the Covid (MSM) hype and those that do not. In fact it’s a 4 way split 😉
Thank you.
A new pedant who calls himself(?) Stephen T, has recently joined UnHerd. He may squeal at your inconsequential, if slightly unusual spelling.
Although guilty of the same ‘crime’, his hypocrisy outshines his caution.
If you use Chrome instead of Internet Explorer that problem goes away.
When I was young and naive I believed that the role of the journalist was fearlessly to report the truth. Now that I am older and slightly more cynical i realise that the role of the journalist is fearlessly to sell newspapers/gather clicks.
What is more likely to attract attention? “16.7M cases globally” or “Out of 7.7BN people only 66k serious or critical cases from 5.7M current global cases”?
Stop press!
70 million British citizens awoke successfully from sleep this morning.
Now I’ll sit back and collect all that lovely clickthrough advertising revenue.
Shock horror stop press! A significant number of British people died overnight!
I bet I get more clicks. I might add LOL but people would find it offensive and ‘ no platform’ me for their lack of comprehension.
You’re too generous, Tom. Most of the time the mistakes are not innocent. At the best, they are careless, in many cases they are deliberate, in pursuit of an agenda.
“Tell me what you want to prove, and I’ll give you the numbers to prove it.”
-Me, in a ~2013 board meeting when I was fed up of getting yelled at for the figures being low.
As an accountant I got the numbers right- just in the wrong order.
It can also be an issue for accountants if they are the wrong colour – i.e. red when they are supposed to be black.
“Lies, damned lies, and statistics”
The only statistic that matters here is that C-19 is a killer of the old and the ‘knackered’ to use the vernacular.
The Swedish figures, so perfectly presented by Elizabeth Hart on another post, on this subject make this abundantly clear.
Perhaps at later date, we may discuss what ‘we’ really mean by Care/Death Homes. So far this topic has been studiously avoided.
Exactly so. Only 20 people under 20 and another 210 under 40 have died with CV in hospital in England, and one suspects most had several co-morbidities.
Total deaths in the UK are now below seasonal average, though more people than usual are dying at home rather than in hospital
Of course this is no consolation to people whose elderly relatives died a few months early, without the comfort of their family with them, but statistics are cruel.
Thank you. Yes, statistics can be cruel, but so can the truth.It was ever thus.
How extraordinary then that the frankly hysterical response of HMG has permanently blighted the lives of so many of our “angry and defrauded young”.
This Government, and virtually every other Government.
With the notable exception of Sweden, you will agree?
“Fools seldom differ”. No doubt by Christmas we will be reaping the whirlwind of this madness.
ONS breaks figures down as
age under 14: 6 deaths to 17 July
15-44: 551
45-64: 4895
65-74: 7549
75-84: 16586
85 up: 21677
But perhaps the most useful way of looking at it is due to David Speigelhalter, who has observed that in all age-groups the proportion dying of Covid this year is roughly the same as the proportion dying of any cause in a normal year.
Thanks. It would be very interesting to see those 45-64 figures broken down.
Quite a lot changes during those vital twenty years, which is probably why the ONS hasn’t produced them?
The figures are available from ONS in great detail, but not compiled in quite the form you want. There is a dataset at Deaths involving COVID-19, England and Wales which shows that for each of the four months March-June, the perdentage of deaths due to Covid in each 5-year age-group from 15-19 upwards was pretty much the same as the percentage of deaths due to Covid across the adult population as a whole. For example, for April
April
England and Wales
Age Number of deathsPercentage of deaths in age-group
PersonsMaleFemalePersonsMaleFemale
29,06416,45112,61336.0%40.1%31.7%
15-19 541 20.8%26.7%11.1%
20-24 1394 25.0%28.1%20.0%
25-29 301911 31.9%31.7%32.4%
30-34 453015 31.5%36.6%24.6%
35-39 814536 30.8%31.3%30.3%
40-44 1419150 31.9%35.1%27.3%
45-49 27018189 32.7%36.5%27.0%
50-54 528328200 35.9%38.4%32.6%
55-59 878610268 37.1%41.1%30.3%
60-64 1,282864418 38.3%41.3%33.4%
65-69 1,6881,137551 37.0%40.1%31.8%
70-74 2,7121,764948 36.6%39.4%32.2%
75-79 3,9182,4361,48238.5%41.2%34.8%
80-84 5,5623,2262,33639.5%42.5%35.9%
85-89 5,8193,1442,67536.0%40.7%31.7%
90+ 6,0912,5633,52831.9%37.7%28.7%
So presumably there are still plenty dying of non-Covid causes in addition to those. Or, to put it another way, the death rate is about double what it normally is.
“in all age-groups the proportion dying of Covid this year is roughly the same as the proportion dying of any cause in a normal year”
I don’t see how that’s correct, unless it includes a forecast that COVID deaths are going to increase by greater than an order of magnitude from where they are currently.
In the U.S., for example, the current total number of deaths attributed to COVID is just over 150,000. By comparison, ~2.8 million people typically die each year in the U.S. Logically speaking, that claim from Speigelhalter says that 2020 COVID deaths would reach the same 2.8 million total.
The math should be similar in the UK (and other countries). COVID deaths in the UK (per Worldometer) are currently ~680 per million people compared to ~470 per million in the U.S., and life expectancy isn’t *that* much different between the two countries.
Avoided for two reasons IMHO. First it would require the “officials” to split the rates and thus allow people to see that other things are far more dangerous if you are under 50. Second reason it would highlight how badly the state and/or NHS and/or care homes were at managing the safety of a known at risk group.
I doubt it would change the mind of some of the sheep who drive cars on their own with a mask….sheesh!!!
The voice of reason. Thank you!
Really hate that quote as it is often used as a big brush to tar any accurate analysis with being misleading.
The issue is that statistics are so often abused, or even good and bad in some cases, and people get duped because they do not scrutinise the figures to the correct level of detail.
Aware that’s not quite the point you were making – but had to rant 🙂
I was rather surprised it hadn’t be used earlier in this discussion.
It has been used for well over a century, so as you imply, it is time to let it rest.
Statistics are information. Like other information they can be accurate (or not), misused (or not), etc. This quotation is trotted out by people whose arguments have been invalidated by statistics in a way that they can’t refute. No one ever says this when the stats are in favour of their stance.
Well put – far clearer than I did
“Lies, damn lies, and sentences of written language.” But of course wordsmiths wouldn’t be so keen on that one.
I get your point, that the stats can be mis-leading but beyond that even, I am not sure the way the stats are collated in the first place can be relied upon anymore. The Govt changed the rules about recording deaths so in what proportion of Covid-19 ‘mentions’ was the disease actually present? And in how many cases, if actually present, was Covid-19 responsible for death? We just don’t know. How many of the excess deaths during the epidemic are due to Covid-19, and how many are due to healthcare reorganisation, lockdown and social distancing? Again, we don’t know.
The media concentrate on Covid cases, counts of which are dubious, over deaths which are somewhat more accurate. Using the two as a ratio gives a wrong estimate of the fatality rate, and is especially bad when uncorrelated to age.
To absolve the media of any culpability for this extremely serious problem is being very generous to the profession of journalism. My concern is that having identified the problem and explained the reasons behind it the writer just accepts that nothing will change and mistakes (genuine or deliberate?) will continue to be made so we must all be on our guard. In my world we would always try to learn from our mistakes to improve the quality of our service for the benefit of all our stakeholders.. Would it not be more constructive if possible solutions were also put forward? Surely all journalists have a responsibility towards CPD ( continual professional development). Therefore, assuming they are aware they are “numerically challenged”, perhaps they should take positive action to try and improve their knowledge and skills in these matters? Likewise their employers could provide training or better still utilise the services of a qualified mathematician/statistician who is responsible for ensuring these mistakes are not made in the first place. For the journalistic profession to just accept the status quo would indicate they have no respect for themselves and their stakeholders, especially their readers…….unless of course their mistakes are disingenuous and deliberate.
Media hype seems to me to be distorting risk both ways (up and down), depending upon which media you read. The maths around Covid-19 spread is really quite tricky; I have rarely seen a media source where the author appears to have a genuine understanding of it and even amongst the scientists, such an understanding seems to be elusive. Don’t get me started on btl commenters and the Twitterati. A great case study is that of the chloroquevangelists, but I’ll leave that for another day.
No, I don’t see it distorted both ways, not to the magnitude of the MSM. My favorite is Newsweek publishing a headline saying that cases spiked after Trump’s rally in Oklahoma two days after the rally when it takes at minimum 4 days for symptoms to manifest. In the body of the article they even admitted that the spike was not due to Trump’s rally but “experts feared it would cause another spike”. They changed the title and the article a couple days later, but the damage (and the purpose) were already done.
So actually you do see it distorted both ways, but consider that the magnitude in one direction is different from that in the other. Then you have one single anectodal example.
Oh, I have TONS of examples, and you’re right. I misspoke. I do not see the right wing distorting covid…not one person has given me a single example where I could not find data to disprove it.
Journalists are awful at presenting numbers. The most usual one of late is to present infection rate for covid when what they actually mean is infection count. There are many other examples, in fact, too many to mention. The worry is that the public are in the main pretty innumerate.
I think they know what they’re doing by and large.
Add to that with a significant proportion of the media online – the journalists are quite literally incentivised to generate views/clicks – so the more dramatic their piece the more it will get shared etc. They are in trouble financially and driven by clicks and views.
That said a little bit of it may also be statistical illiteracy – most journalists do not come from a STEM background and so will not have the required statistical discipline built in.
“The Covid19 epidemic has forced everyone to undertake a rapid and high-stakes crash coure in statistics. Suddenly, everyone, including journalists, has to be conversant in things like Bayes’ theorem, statistical modelling and Goodhart’s Law”
Eh? I doubt if few people, least of all journalists, have the foggiest idea about statistics and how to interpret them.
What the writer avoids saying is that the mainstream media deliberately, as a policy choice, makes what ever conclusions it likes from its selected statistics. It is not ignorance per se which leads them down these dark alleys, but their own conscious determination to make everything seem dangerous and a consequence of doctrinaire government decisions. Now if only the MSM were in control, all these problems would be avoided – well, they wouldn’t report ’em, would they?
Believe the author is letting the media off the hook some by calling their distortions innocent. I think in one sense they are: The journalists probably believe what they are writing is “true.” But if they have a particular ideological bent, the editors will tend to select in the first place things that happened in the world yesterday that conform to this leaning as “representative” and newsworthy and overlook the many other things that may contradict their narrative. It may be unconscious, but it is not innocent.
Agree fully, it’s another example of a broader problem….there is an increasing number of self researched ‘experts’ on all matters.
Statisticians have a professional body with minimum standards for membership however the cross over between that and the statistical modellers of the scientific community seems to be an unexplored grey area…
Correct. It is all about the denominator for all of these stories. But now it is most important to know of the Covid deaths as a numerator over the entire population and to see age and other factors in the stats. Then we can all make intelligent decisions .
How far would education need to go in teaching statistical methods to render the public at large able to sift the speculative from the replicable stats? Probablyend of 1st year exams for STEM undergrads is my untested guess. So we are both a long way off and heading in the wrong direction. Journalists appear on average way behind the public at large in honesty as well as intelligence. By trumpeting the wild theories of Imperial College, UN Rapporteurs, Marmots, etc etc they are making things worse for society as a whole, them included. If nothing is done to balance this onslaught they will become like journalists in Mexico now or Brazil/Arentina in the ’70s. People quite logically exaggerate qualatitive risks ahead of quantative ones. Effect of journalists, left, right and middle egging people on into hysteria and civil unrest = burning, looting, murder. Chances of us being murdered in such conditions = quantatively low, but in risk analysis being murdered is a Terminal Event. Therefore it makes sense to stamp the risk out at source. Mr Journalist – meet Mr Pararamilitary Death Squad. It seems the media have no idea of the path they have embarked on. Lets hope they figure it out or some grown ups arrive and stop the party before it gets out of bounds.
If you’ve been reading the NYT Daily Brief the last few months, you’ll note that every single brief starts with some kind of woeful news or statistics about the virus. I believe the writers are following standard marching orders: “keep the virus news front and center and make sure it’s all bad news”. If I can make it all the way through a NYT brief I usually feel like strangling myself with my own colon. But I suppose that’s the point.
Maybe using the NYT is using the dependent variable.
The N.Y. Times historically the good standard for journalism has changed their strategy to cater to a specific profitable market segment. As a result too many of their pieces include unsupported opinion and even misuse of the dependent variable or unrelated facts.
Had to dissect a couple of pieces they wrote recently: The truth about how media is misleading with facts
Poor Hillary, I’m in tears. At first I thought the writer was being sarcastic…Who cares if someone is a liar, manipulative and power-hungry, besides being a friend of Jeffrey Epstein and flying to his “orgy Island” at least six times? There have been claims of the orphanage the Clintons ran in Haiti being used to traffic children to the US. I just can’t stand people like this author who excuse anything just because the person is a woman and constantly play the victim card. What matters is a person’s character, not their sex.
I have always tried to focus on the personal risk to myself and the stats that are quoted in the press are not much help. For example, my understanding is that the risk of catching the virus when social distancing at 2 metres is 1.3% which doubles, still only another 1.3%, to 2.6% at 1 metre. Does this mean that the risk of not catching the virus is 98.7% at 2m and 97.4% at 1 metre? Why not focus on the risk of not catching rather than catching the virus.
Exactly – in my area we have a spike in cases apparently (obviously nothing to do with increased testing), How many cases? The reference population is 100,000, the spike being 21 cases – 0.021%. What’s the chance of me being infected? What’s the chance of my experiencing mild symptoms? What’s the statistical chance of my being severely ill? Media talk of locking down. What’s my chance of getting cancer? One in two of us will apparently get cancer. How many of those will die, especially if there’s no early diagnosis? What are my chances of dying of cancer against covid?
It’s not so new that the media has difficulty with statistics. It is new that when statistical facts are presented that refute media beliefs, they continue to use misleading information. Such misinformation is no longer uninformed, rather it crosses the line to propaganda.
When every piece is crafted from a political perspective it’s unhealthy for us all. Uniformed or misinformed population and leaders will result in disastrous policy decisions.
I think you are being far too generous with “None of this, we want to stress, is deliberate, or even especially blameworthy; journalists make mistakes, and like the rest of us, most are not especially numerate.”
When the auto-cue reader, who is paid a small fortune for reading from the auto-cue, turns to the special correspondent on the subject matter, then that special correspondent ought to be competent and have done his/her homework. The problem is the truth is not very interesting, but misrepresentation which suggests something which probably is not true in support of some agenda is far more interesting to those pushing the agenda. These are not blameless mistakes.
I switched off BBC news in disgust yet again when they reported the figures on who had been fined for breaching lock down. The first thing was more young people were fined than old people – not surprising and that was the comment made by the auto-cue reader and probably is true and can be taken at face value. Then the percentages of the White population, the Black population and the Asian population were given, showing a clear bias towards fining Blacks and even more so Asians. The special correspondent said it might be because these people don’t trust the police and escalate the situation when the police start on the 4 Es. Hang on a minute! the simple fact is the proportion of Black and Asian who are young in this country is far greater than the proportion who are young in the white population – the median ages differ by over 10 years. So the data needs to be appropriately adjusted for that fact before you start to compare and draw conclusions as to possible cause – at least this time is was not portrayed as evidence of police racism, but I bet there were a lot of people who think the BBC were just making excuses for police racism.
Far too generous indeed. If your ‘profession’ is reporting on matters to the public, it is simply not acceptable to shrug your shoulders and say you can’t add up. The inadequacy of the media in this area is shocking. But then, they don’t appear these days to be able to spell or understand grammar which should be their stock in trade, so perhaps numeracy is a bit much to expect.
Whatever the ostensible subject of Tanya Gold’s articles, there’s always an undercurrent of self-pity and misandry. An excellent rule is to discover what side she’s on, and then settle for the opposite.
I strongly disagree. In fact every one of us has a Cliffordian moral obligation but journalists are the (necessary) intermediaries between “science”, broadly understood, and society and have a far greater responsibility to uphold it. Yet they frequently misinform us (and in many ways – not just with ‘small’ factual errors) and routinely neither acknowledge nor correct their mistakes. Not even their honest ones.
Mathematical modelling is useful insofar as the parameters are pertinent, some of the early Imperial ones weren’t sufficiently up to date with what was then becoming known about this virus and its crucial differences from SARS and flu in both infection and pathologies. When these were updated there were massive changes in the projections.
Just now it’s becoming evident that the higher and more diverse social mixing associated with holidaymaking is incompatible with the identify and control strategies necessary to control clusters and avoid second general lockdowns. The chains of events that arise are easily subject to mathematical modelling and it’s hard to believe governments weren’t given projections. The Robert Koch Institute has posted on this, although only recently. Why has the boat been sailed right up to the rocks before trying to change course?
Someone with as much data savvy as Mr Cummings must have seen the contradiction between control strategies to identify and separate the infected and increased number , range and randomness of social encounters, whatever the distancing.
‘In a subtle way, though, this is what the media does all the time. Like the water-causing-death study, it selects on the dependent variable: and for the media, the dependent variable is whether it’s interesting.’
There has been nothing subtle about this. And it is one of the reasons so many of us gave up on the MSM many years ago. As for the whole causation/correlation thing, I once pitched a programme called ‘Correlation Street’ in which each episode would feature an entertaining example of correlation not being causation. Sadly, it was not commissioned.
What are the numbers measuring and are they accurate.
For example – how many who died with Covid 19 on the certificate died from only that? Have we measured the covid deaths consistently with how we measure flu deaths?
We see a rising infection rate – how many receive treatment or more importantly how many new covid beds are being taken up in hospitals – this is what we shut down for to save the NHS being overrun with covid cases.
How many new covid cases requiring hospitalisation are going to nightingales or covid limited hospitals?
Maths and stats are as much about the question as the numbers.
When Mr Whitty ( tempting to nickname him DimWhit ) provides advice it always seems to me to be too focussed with no context or big picture. For example 10 people come back from Spain and test positive – well if we had a compulsory test, track and trace app and system in place those people couldn’t only be placed in a hotel (or some of Londons vacanct office space…) for 14 days but also all those Brits they mingled with on holiday could be TTT’d
you would probably find more positive std tests among returning youth than covid 19 and do we anything about that?
Great, thankyou all.
Little boy with sore finger says to Doc Martin “Am I going to die” Doc Martin says very dismissively “Of COURSE you are, ..everybody dies”. Oh yes love Doc Martin & UnHerd , Cheers from Australia.
http://www.ronpaullibertyre… Ron Paul “From Sweden To Uruguay The Evidence Is Clear: Lockdown ‘Cure’ Worse Than The Disease”
It’s the same in do many successful companies. Often one, but sometimes two or three driven individuals who make things happen. Of course with bands and companies the customer can walk away if they want to. That’s why governments can not be dictatorships.
The media demand attention, that is their entire business model, whether it’s the BBC or Daily Mail. They will present a story in the most compelling way to attract that attention. Their baseline of truth is simply to avoid a lawsuit and therefore are free to distort, twist and select data to suit that need to encourage people to watch the show, buy the paper or click the link. It will always be like this – sadly it’s a facet of human nature.
I think Peter has identified one of the possible causes of the journalistic world’s failure to inform and enlighten us.
But journalists young and old are all obsessed with finding a story that is “newsworthy” and the more unlikely and ridiculous the better.
The PCC has 7/17 top slots occupied by editors which means that only 2/10 non journos need to vote with them to give a majority in favour of the continued abuse of press freedom. Also the PCC is funded by the media and hasn’t even begun to cope with a world with social media playing a major part.
I think the public are more savvy as a result of the media’s outrageous coverage of covid. My wife and I have banished the worst culprits from our repertoire forever with BBC Radio 4 Today and all BBC TV news removed from our schedule first.
Others have followed – many female journalists included too unfortunately.
Thank you again Tom and Unherd for the stimulating antidote
Correct about the biases. Incorrect on two points: The men going into shelters “claiming” to be trans ARE trans and not simply claiming they are. Transgender is anyone pretending to be the opposite sex via external adoption of sex based stereotypes of the opposite SEX or via surgeries that imitate visually the sex characteristics of the opposite sex and taking unnatural hormone doses of the opposite sex and pills to decrease the naturally occurring sex hormones. The claim they are “pretending” is meaningless since no matter what end of the spectrum here, they are ALL pretending and many who’ve been pretending to be the opposite sex for MANY YEARS go into women’s shelters, bathrooms and prisons to rape and abuse women. Full stop.
Second: You insinuate COVID-19 IS more dangerous than the flu w/no evidence for that assertion. Evidence DOES however show this is NOT the case in the majority of the population when stratified by age and severe underlying health conditions. The “death count” is also inflated by several countries listing all deaths where a “positive” (that itself is entirely another discussion) test *may* show up but the causal factor was something else. Furthermore, influenza is under-counted and this has been a problem for many years. It also more importantly has a vaccine. Looking at the numbers by country, one does not need more than simple math skills to appreciate the astronomical differences between countries. And then the next question should be WHY is that so.
Those points have nothing to do with mathematical bias and are an entirely different form of reporting bias.
It is not just media that distorts risk. Government also grossly distorts risk to attack civil liberties, the human right of privacy and free speech. Both Tory and Labour governments have grossly exaggerated the risks of terrorism to institute a system of surveillance that would make the Stasi blush. At one point GCHQ was illegally downloading millions of private videos on yahoo.com. Why? You don’t combat terrorism by grossly infringing the privacy of ordinary law abiding citizens. Right wingers seem particularly keen to exaggerate the risks from Islamic terrorism. I wonder how many of them also find it expedient to dismiss the risks of covid?
Objective informed journalism (as here) has been systematically removed from the mainstream – this much is clear – and the mainstream consumers have not the time or wit to find ‘the truth’ for themselves. No journalist with any integrity should release articles that they have not fully researched or are not qualified to write. Personal integrity seems to be at an all time low in the media and there seems to be any number of them willing to compromise to have a job and ‘toe a line’ or more odiously to feed some callow ego need. Swiss Propagada research has some scary research pointing towards the US exhibiting the highest level of spin/propaganda on the planet. As Carl Jung said – the search for truth and consciousness is a hard, lifelong job of work – and few choose to undertake it (paraphrased). Tragically it seems that that journey is not the road most travelled and few have even reached Socrates first step towards wisdom – to realize that you know nothing – leading to his next 4000 steps which are to question everything……
Government is a particularly guilty party. Testing is a process that has a denominator (the number of tests) as well as a numerator (the positives). The UK Government has done everything it can to obscure the relationship. This has extended from not publishing local data on the numbers of tests conducted in “hotspots” to not giving any comment at all on the rate of positive tests nationally. You would never guess, in a week when Government has been warning about a 28 per cent jump in cases (and taking new Draconian steps in the North of England) that the ratio is half what it was at the end of June, and has been stable between 0.005 and 0.006 for nearly a month.
Stats fans might be interested too in the “Doctoring Data” book by Dr. Malcolm Kendrick, particularly in the current Covid climate.
I saw a wonderful video of a lecture of his recently where he showed how a pharmaceutical can conjure up a 35% improvement from data of virtually zero.
His writings are quite dry and full of facts, but that appeals to the geek in me 😉
So, she apparently ironed Bills shirts in the full and certain knowledge that one day she would be a Secretary of State. Yeah right. I was certain my dad would be Prime Minister when I once shined his shoes for him.
I very much doubt that she ever ironed her own clothes since leaving college.
Why do these wealthy from birth types claim rights to power on the back of that while stealing the clothes of the poor and powerless to win arguments?
Tom, excellent article, and a point well made.
Despite the fact that you might see me as competition, please have a look at my Book “Good Thinking & Bad” (on amazon etc). It is a slightly wider scope and conceived as a text book to provide thinking tools, but you might find some useful references to support your views.
This is far too easy on journalists. Their public service requires a good basic education and a lack of rudimentary maths falls well below that.
A good indication of a person’s character is how well they treat those who look after them; Mrs Thatcher was kind and thoughtful to all the Downing Street staff, and always personally thanked the cooks and waiters after a formal dinner. I would be interested to know whether Hillary Clinton was a pleasure to work for.
Sometimes I cannot believe the premium content I’m getting here for free. Well written, objective and shareable. Great job and thank you for that!
Must watch!
Americas Frontline Doctors Address COVID-19 Misinformation & Hydroxychloroquine truth – Capitol Hill Press Conference
https://videos.utahgunexcha…
Why has hydroxychloroquine been disregarded?
First, as all know, the medication has become highly politicized. For many, it is viewed as a marker of political identity, on both sides of the political spectrum. Nobody needs me to remind them that this is not how medicine should proceed. We must judge this medication strictly on the science. When doctors graduate from medical school, they formally promise to make the health and life of the patient their first consideration, without biases of race, religion, nationality, social standing”or political affiliation. Lives must come first.
https://www.newsweek.com/ke…